We propose to investigate the cellular properties and systems-level mechanisms used by a simple nervous system to make decisions about selecting behaviors. In the past 27 years, my laboratory has helped to characterize the neuronal circuits in the nervous system of the medicinal leech that are responsible for four different behaviors: local bending, whole-body shortening, swimming, and crawling. In addition, we have found a number of neurons that are involved in activating and terminating these behaviors. In recent years, we have helped to develop a new generation of voltage-sensitive dyes that are so much more sensitive that we can now record the electrical activity of a many neurons at once. With parallel improvements in the sensitivity of CCD cameras along with faster computers with huge memories, we can now realize what was until very recently just a fond hope: to record from all the relevant neurons in a leech nervous system as the animal makes decisions among behaviors. Our previous electrophysiological data indicated that decision-making neurons are not dedicated to the initiation of a single behavior, but instead help to initiate two or more behaviors, some of them mutually incompatible (e.g., shortening and swimming). These kinds of findings indicate that decisions are made by a combinatorial code, in which a cluster of decision-making neurons determines which behavior is produced. We will determine the nature of this code, and will test the importance of each of the neurons in producing the ultimate decision. In particular, we will test whether all decision-makers have equal influence or whether instead, whether the decision is weighted among neuons, and whether decisions are made hierarchically, with general decisions being made before more specific ones. In addition, we will use the tools of nonlinear dynamics to help us to understand the code, and to capture the essential features of the decision-making process. Also, we will use our knowledge of the mechanisms of decision-making to determine where in the nervous system changes take place that turn one response into another when an animal learns a complex task. The kinds of interactions being posited for decision-making in the leech have been postulated in more complex nervous systems, including our own. Because the leech nervous system is relatively simple and very accessible, and because of the technological advances described above, we are now in a position to test these ideas on a neuron-by-neuron basis.